Patient-Reported Barriers and Enablers to Deprescribing Recommendations During a Clinical Trial (Shed-MEDS)

Abstract

Background and Objectives

Effective deprescribing requires shared decision making between a patient and their clinician, and should be used when implementing evidence-based deprescribing conversations. As part of the Shed-MEDS clinical trial, this study assessed barriers and enablers that influence patient decision making in deprescribing to inform future implementation efforts and adaptations.

Research Design and Methods

Shed-MEDS, a randomized controlled deprescribing trial, included hospitalized older adults discharging to post-acute care facilities. A trained clinician reviewed each participant’s medical history and medication list to identify medications with potential for deprescribing. The study clinician then conducted a semistructured patient-centered deprescribing interview to determine patient (or surrogate) concerns about medications and willingness to deprescribe. Reeve et al.’s (2013) framework was used to categorize barriers and enablers to deprescribing from the patient’s perspective, including “appropriateness of cessation,” “fear,” “dislike of a medication,” “influences,” and “process of cessation.”

Results

Overall, participants/surrogates (N = 177) agreed with 63% (883 total medications) of the study clinician’s deprescribing recommendations. Thematic analysis revealed that “appropriateness” of a medication was the most common barrier (88.2%) and enabler (67.3%) to deprescribing. Other deprescribing enablers were in the following domains: “influences” (22.7%), “process” (22.5%), “pragmatic” (19.4%), and “dislike” (5.3%).

Discussion and Implications

Use of a semistructured deprescribing interview conversation tool allowed study clinicians to elicit individual barriers and enablers to deprescribing from the patient’s perspective. Participants in this study expressed more agreement than disagreement with study clinicians’ deprescribing recommendations. These results should inform future implementation efforts that incorporate a patient-centered framework during deprescribing conversations.

Clinical Trials Registration Number

NCT02979353

Polypharmacy, typically defined as concomitant use of five or more medications, affects 40%–50% of older adults in the United States and is associated with geriatric syndromes, medication nonadherence, adverse drug events, and poor health outcomes (Kane et al., 2012; Saraf et al., 2016). While polypharmacy may be appropriate among those with multiple chronic medical conditions, a lack of clinician reassessment for appropriateness or effectiveness can lead to potentially inappropriate medication (PIM) use (Zullo et al., 2018).

Deprescribing, the systematic process of medication cessation, has been identified as an effective way to mitigate polypharmacy and inappropriate medication use (Linsky et al., 2019). Recent studies suggest that the majority of adults aged 65 and older express a willingness to deprescribe a medication if their clinician agrees, and that they wish to be involved in the decision-making process (Chock et al., 2021; Garfinkel et al., 2010; Hollingsworth et al., 2022; Reeve et al., 2013). Several tools have been developed to ascertain patients’ attitudes and willingness to engage in deprescribing, yet these tools have been shown to have poor predictive validity and do not fully capture deprescribing barriers and enablers from the patient’s perspective. Furthermore, patients’ attitudes toward their medications and willingness to deprescribe often vary by medication (Weir et al., 2018). To understand medication-specific barriers and enablers, clinicians need to engage patients in deprescribing conversations to ensure that their preferences, goals of care, and concerns are known (Linsky et al., 2017, 2020; Reeve et al., 2013, 2016).

Effective deprescribing requires shared decision making between a patient and their clinician, in that decisions should be influenced by factors unique to the individual such as medication cost, the experience of side effects, and goals of care and/or life expectancy (Le Bosquet et al., 2019; Pickering et al., 2020; Scott et al., 2015; Thompson et al., 2019). A patient-centered clinician–patient conversation solicits barriers to and enablers of medication cessation that help a clinician understand how patients’ unique factors contribute to their willingness to deprescribe. Thus, the ability of clinicians to implement deprescribing successfully relies on eliciting the barriers and enablers that potentially influence patient decision making related to medication use (Gnjidic et al., 2012; Page et al., 2016; Potter et al., 2016; Reeve et al., 2018). The Reeve et al. (2013) framework categorizes barriers and enablers to deprescribing from the patient’s perspective, thereby supporting a patient-centered deprescribing process.

The importance of shared decision making in the deprescribing process has been consistently identified as a critical component of effective deprescribing (Chock et al., 2021; Jansen et al., 2016; Turner et al., 2018; Weir et al., 2018). Evidence-based deprescribing guidelines and algorithms recommend patient-centered approaches and endorse the integration of shared decision making (Bruyere Research Institute, n.d.; Reeve et al., 2019). Yet, few studies describe the process and/or content of deprescribing conversations between patients and clinicians in a manner that would allow replication in clinical practice.

The Shed-MEDS randomized controlled trial (NCT02979353) was a deprescribing study among hospitalized older (≥50 years) adults discharged to post-acute care (PAC) facilities (Vasilevskis et al., 2019). The primary aim was to determine if a deprescribing intervention initiated during hospitalization and continued through the PAC stay reduced the total number of medications among older patients when compared to usual care. We utilized a patient-centered deprescribing conversation framework to support shared decision making when making deprescribing recommendations. The purpose of this paper is twofold: (a) to describe the process for facilitating the patient–clinician deprescribing conversation to support shared decision making, and (b) to summarize the most common barriers and enablers expressed by older hospitalized patients in response to specific deprescribing recommendations presented during the patient–clinician conversation.

Method

The Vanderbilt Institutional Review Board and the National Institute on Aging-appointed Data Safety and Monitoring Board approved all study procedures. Inclusion criteria for the larger clinical trial were hospitalized adults aged 50 and older, referred to a PAC facility, and taking five or more medications prior to hospitalization. Those with cognitive impairment were eligible for the trial if they had surrogate consent. The published study protocol outlines the design of the parent study (Vasilevskis et al., 2019).

Figure 1 provides a summary of the intervention protocol steps, which includes the patient–clinician deprescribing conversation process. As part of the Shed-MEDS intervention, a trained study clinician, either a pharmacist or nurse practitioner (NP), conducted a detailed medication history and reconciliation for all study participants to yield a Best Possible Medication History (Shah et al., 2021). This list included prescription and over-the-counter (OTC) medications, for both scheduled and as-needed use. Those randomized to the intervention group had their medication list further reviewed by a study clinician to identify medications with the potential for deprescribing, defined as dose reduction, titration, stopping of a medication during hospitalization, or recommending a time frame for the medication to be stopped following hospital discharge. This clinician prereview process included an extensive medical record review to yield a pairing of indication(s) to each medication, recommendations for continuation or deprescribing based on indication(s) and evidence-based recommendations, and identification of one or more clinical rationales for each deprescribing recommendation (Scott et al., 2012). Study protocols related to the medication history and prereview processes have been published previously (Shah et al., 2021; Vasilevskis et al., 2019).

Figure 1.

Patient–clinician deprescribing conversation process.

Patient–Clinician Conversation About Deprescribing Recommendations

After completing the medication prereview process, the pharmacist or NP prepared the structured patient–clinician conversation tool with the targeted list of prehospital medications identified for potential deprescribing (see Supplementary Material). The clinician prioritized medications with a higher risk of harm or potential harm, according to the Beer’s Criteria (American Geriatrics Society, 2015), Screening Tool of Older Persons’ Prescriptions (STOPP) (Gallagher & O’Mahony, 2008; Gallagher et al., 2009), and Rationalization of home medication by an Adjusted STOPP list in older Patients (RASP) (Van der Linden et al., 2014) lists.

The patient–clinician conversation included a semistructured interview conducted at the hospital bedside with the patient, their surrogate, or both. Surrogates were included in the conversation if the patient’s cognitive deficits precluded their ability to fully participate, if the patient preferred to have the surrogate involved in decision making, or if the surrogate had a role in the patient’s medication management. Structured questions within the conversation tool (Supplementary Material, Parts A and B) were adapted from the Brief Medication Questionnaire (BMQ; Svarstad et al., 1999). The study clinicians took written field notes during the entire patient–clinician conversation, including the patient’s general comments about medications and prescribers, in addition to their reasons for wanting to continue or deprescribe each specific medication.

The first section of the conversation tool, General Medication Management-Related Issues (Part A), included a general evaluation of problems or concerns regarding the patient’s medication regimen from the patient’s perspective. Patient (or surrogate) responses to these questions informed the conversation during the second part of the interview, Medications Targeted for Deprescribing (Part B), which focused on each medication recommended for deprescribing by the study clinician in addition to those identified as a concern by the patient or surrogate.

Patient (and/or surrogate) responses to both Parts A and B often led to the clinician providing education about a specific medication prior to discussing the rationale(s) for deprescribing. Education was tailored to patient and/or surrogate concerns or knowledge deficit and included information about the medication’s indication(s), administration directions, common side effects, and potential benefits and risks of harm, based on the available evidence. For example, the study clinician may have addressed a patient’s misunderstanding about a medication’s indication (patient’s understanding differs from the actual indication) and provided education about associated side effects or health risks (e.g., falls). Throughout the deprescribing conversation, the clinician addressed patient or surrogate concerns about medications and deprescribing, and together, a plan was created for each medication with deprescribing potential.

Categorizing Patient Responses Into Barriers and Enablers

The conversation tool (Supplementary Material) was utilized to document the outcome of the patient–clinician conversation. When a patient (or surrogate) was not agreeable to the deprescribing recommendation, the plan included the continuation of the medication, and the clinician documented the reason(s) provided by the patient (or surrogate) for disagreement. When the patient was agreeable to deprescribing, the clinician documented the reason(s) provided by the patient (or surrogate) for their willingness to deprescribe. A deductive thematic analysis was conducted of the patient’s (and/or surrogate’s) expressed barriers and enablers to deprescribing, based on categories defined in Reeve et al.’s systematic review and deprescribing framework (2013), which included appropriateness, process, influences, pragmatic, fear, and dislike (Table 1).

Table 1.

Patient Barrier and Enabler Categories in Response to Deprescribing Recommendations

Category	Category definitionsa	Shed-MEDS examples
Barriers
Appropriateness	Believes that a medication is an important part of the patient’s medical care Trusts that the medication has continued or future benefits Disbelief that the medication is causing any harm	“[The Patient is] not willing to stop; [he] had a flare when he neglected taking [it] one time.” “[She] has taken [the medication every day] for a long time; it controls her allergy symptoms and [she] wants to continue [taking it].” “[The patient believes that] this is the only thing that helps.” “[The patient states] ‘I need this.’ [She] takes scheduled Miralax in addition to other bowel agents and reports a [bowel movement] frequency of [every] 3 days.” “[She] doesn’t feel like [the medication is] harming her and that it has helped with her [gastrointestinal] issues overall.”
Fear	Concern about new or recurrent symptoms that follow deprescribing Fear that medication cessation will cause withdrawal symptoms, death, or a return of the medical condition for which the medication was prescribed	“[She] fears having a stroke; [she] started [the medication] 10 years ago. [She] doesn’t know if it helps, but she hasn’t had a stroke.” “[The] patient said it’s working very well and she finally feels ‘stable.’ She’s afraid her symptoms and mood will get worse if she stops [the medication] and she’s worked so hard to get here.”
Process	Frustration that the deprescribing process is complex Belief that deprescribing is better suited for a different time	“Let’s wait until I’m more stable to change this [medication].” “[The patient’s] daughter doesn’t want to change this [medication] because she wants to wait to see what [the] patient’s baseline mood is after hospitalization.”
Influences	Negative previous experiences with stopping a medication Prior discussions with a medical provider about the medication’s importance	“[He] previously tried to stop [the medication] but [his] symptoms returned. [He] wants to keep [it].” “[She] will continue [taking the medication] because her cardiologist says she needs it.” “My doctor says [he] doesn’t like anyone [to] mess with my propranolol.”
Pragmatic	Long time use of a medication Patient is resistant to change Assumption that the medication is a safer option than other alternatives	“She has been taking [the medication] for almost 30 years. She does want to reduce [the dosage] eventually, but not right now.” “[The patient states], ‘It’s not as bad as a statin.’”
Enablers
Appropriateness	Doubts the medication’s necessity or Has been influenced by negative past experiences with the medication	“The patient has had episodes of really low blood sugar.” “[He] has been taking [the medication] ‘a long time’; [he] doesn’t know if it helps or not.”
Dislike	Has experienced specific side effects related to the medication General dislike of taking medications	“[She] states that [the medication] inflames her throat and nasal passages; [she] doesn’t like it.” “[The patient states], ‘[I take] too many supplements. I want to narrow them down to what really works.’”
Process	Knowledge that a medication can be restarted as needed Participation in a deprescribing process that includes clinician follow-up support	“[The patient’s] daughter agrees to monitor and restart [the medication] at lower dose if it is needed.” “[The patient’s family agrees to deprescribe the antihypertensive] and agrees for SHED-Meds to monitor … [the dosage] in hospital and at the [skilled nursing facility].”
Influences	Clinician endorsement of medication cessation Exposure to information from outside sources Previous positive experience(s) with medication cessation	“[The patient states that she is agreeable to stopping the medication] as long as Dr. F says it’s ok.” “[The patient] read about risks [of the medication] on internet.”
Pragmatic	Cost of a medication influences the patient’s choice to favor deprescribing Has already made a previous independent choice to stop taking the medication Concerns about the medication’s compatibility with other medications that are being taken	“[He] isn’t taking [this insulin] due to high cost. [He] wants to continue insulin 70/30, [which is available] at [a] lower cost.” “[The] daughter stated that the patient has already stopped taking this.” “[The] patient reported that this medication just kept being continued and she’s not sure [that] any of her providers have really ‘looked’ at it. In general, [she] reported that she has multiple providers prescribing without really looking at her whole [medication] list.”

^aFrom Reeve et al. (2013).

Following completion of the interview, the study clinician coded patient comments into these predefined barrier and enabler categories. A second study team member independently reviewed the interview comments and coded for agreement. If there was disagreement related to the assigned category, the study team discussed the theme assignments until consensus was reached. The categories were not mutually exclusive in that patient or surrogate comments may have aligned with more than one category for an individual medication.

Analysis

We characterized the patient population with descriptive statistics using means with standard deviations (SDs) and proportions where appropriate. Patient and surrogate responses to semistructured interview questions were reported as medians and interquartile ranges (IQRs). Each deprescribing rationale assigned by Shed-MEDS clinicians during the prereview process was totaled across all medications for all intervention participants. Each medication was classified by the primary indication for use (e.g., depression). We calculated the total number of medications discussed during the patient–clinician deprescribing conversation at the patient level, across all patients, and by medication category. Percent patient/surrogate agreement to deprescribing recommendations was calculated at the patient, medication category, and aggregate levels. We totaled each barrier and enabler category overall and within each medication category.

Results

In total, 372 patients enrolled in the Shed-MEDS trial, 186 of which were randomized to the intervention group. Of those randomized to intervention, 177 (95.2%) completed the patient–clinician deprescribing conversation component of the intervention and were included in the thematic analysis for barriers and enablers. The remaining nine (4.8%) had incomplete patient–clinician deprescribing conversations due to the following reasons: four patients refused the conversation, four patients withdrew consent, and one patient discharged home prior to the conversation.

Table 2 shows the characteristics of intervention participants with complete data (N = 177). Participants had an average age of 76.5 years and were predominantly female (64.4%) and Caucasian (84.7%), with a mean of 16.7 (SD ±6.2) prehospital medications at the time of study enrollment, and a mean of 2.8 outpatient providers. Participants had an average Charlson Comorbidity Index score of 7.1 based on International Classification of Diseases 10th Revision codes, indicative of multiple comorbid conditions, and 27% had a Brief Interview for Mental Status total score of less than 13, indicative of moderate to severe cognitive impairment (Charlson et al., 1987; Chodosh et al., 2008; Saliba et al., 2012; Vasilevskis et al., 2019).

Table 2.

Participant Characteristics (N = 177)

Demographics	Mean (±SD)	% (n)
Age	76.5 (±10.1)
Percent female	64.4 (114)
Percent Caucasian	84.7 (150)
Cognitive impairment (BIMS <13)a	27.2 (43)b
Charlson Comorbidity Indexc	7.1 (±2.9)
Number of outpatient providers	2.8 (1.4)
Number of prehospital medicationsd	16.7 (±6.2)

Notes: BIMS total score range 0–15, with scores below 13 indicative of moderate to severe impairment; SD = standard deviation.

^aBIMS = Brief Interview for Mental Status.

^b N = 158. BIMS evaluation was not conducted on participants with positive Brief Confusion Assessment Method score, indicating delirium.

^cComorbidity-based measure to predict mortality within 1 year of hospitalization; higher scores indicate higher risk of mortality.

^dPrehospital medications included scheduled, as-needed and over-the-counter medications.

Based on the clinician prereview process, a total of 1,710 medications across the 177 intervention group participants were identified as potential targets for deprescribing based on one or more clinical rationales. The most common clinical rationales for deprescribing were No indication for medication (39.2%), Medication is on the STOPP/Beers/RASP lists (25%), and Patient reports not taking medication (15.7%). Additional rationales included Duplicate medication for the same indication (12.9%), Drug–disease interaction (12.7%), and Currently indicated but only for a limited time (7.4%).

Patient–Clinician Conversations

The patient–clinician conversations related to deprescribing recommendations were led predominately by study pharmacists (71.8%), with the remainder led by study NPs. Most conversations were conducted solely with patients (68.9%) while surrogates participated in 31.1% (surrogate only 16.4%, surrogate and patient together 14.7%). The median amount of time required per patient–clinician conversation was 30 min (IQR: 20, 40). Within the General Medication Management-Related Issues section of the patient–clinician conversation tool, the most common areas in which participants reported experiencing difficulty were with side effects (38.4%), opening containers (34.4%), reading the print on packaging (29.9%), and cost (22.5%).

Of 1,710 total medications identified as potential targets for deprescribing, a total of 1,405 (82.4%) medications were discussed as part of the deprescribing conversation, for a median of 9 (IQR: 6, 11) medications discussed per patient (Table 3). Study clinicians were unable to discuss all targeted medications with each patient due to either patient fatigue or insufficient time prior to hospital discharge. The medications discussed were most commonly from the Vitamins and Herbal Supplements (N = 285), and Diabetes (N = 269) medication categories.

Table 3.

Medications (N = 1,405) Discussed for Deprescribing During the Patient–Clinician Deprescribing Conversation and Patient/Surrogate Agreement by Category

Medication category	N of medications discussed for deprescribing	N (%) of medications patients/surrogate agreed to deprescribea
Overall	1,405	883 (62.8)
Anticholinergic	87	52 (59.8)
Antipsychotic	12	11 (91.7)
Anxiety	39	32 (82.1)
Depression	56	27 (48.2)
Diabetes	269	193 (71.7)
GERD: PPI and H2 blockers	63	47 (74.6)
Antihypertensives	101	75 (74.3)
Pain: nonopioid analgesicsb	47	32 (68.1)
Pain: opioid agonists	31	19 (61.3)
Seizure and neuropathic pain	22	14 (63.6)
Vitamins and herbal supplements	285	155 (54.4)
Other	393	226 (57.5)

Notes: GERD = gastroesophageal reflux disease; H2 = histamine 2; PPI = proton pump inhibitors.

^aAgreed to deprescribe is defined as agreement to reduce and/or stopping the medication.

^bExcludes aspirin.

Barriers and Enablers

Overall, patients/surrogates agreed with 62.8% (883 total medications) of the study clinicians’ deprescribing recommendations (Table 3). This translated to patients agreeing to deprescribe a median of 5 (IQR: 3, 7) medications per patient. Patients were most agreeable to deprescribing medications in the Antipsychotic (91.7%), Gastroesophageal Reflux Disease (GERD): Proton Pump Inhibitors and Histamine 2 blockers (74.6%), Antihypertensive (74.3%), and Diabetes (71.7%) medication categories, but were least agreeable to deprescribing Vitamins and Herbal Supplements (54.4%) and medications for Depression (48.2%).

Tables 4 and 5 display barriers and enablers to deprescribing across all medications and by key disease indication medication categories. Overall, “appropriateness” of stopping a medication was the most common barrier (88.2%) and enabler (67.3%) (Tables 4 and 5). Other deprescribing barriers in descending order were in the “fear” (18.4%), “influences” (17.1%), “pragmatic” (7.3%), and “processes” (1.1%) domains. The other enablers to deprescribing fell into the “influences” (22.7%), “processes” (22.5 %), “pragmatic” (19.4 %), and “dislike” (5.3%) domains.

Table 4.

Most Common Barriers by Medication Class

Medication category	Appropriateness	Influences	Process	Pragmatic	Fear
Overall	88% (411)	17% (80)	1% (5)	7% (34)	18% (86)
Anticholinergic	97% (29)	7% (2)	0% (0)	0% (0)	23% (7)
Antipsychotic	0% (0)	0% (0)	0% (0)	0% (0)	100% (1)
Anxiety	100% (6)	33% (2)	0% (0)	50% (3)	33% (2)
Depression	89% (25)	4% (1)	4% (1)	7% (2)	11% (3)
Diabetes	83% (52)	19% (12)	0% (0)	14% (9)	19% (12)
GERD: PPI and H2 blocker	81% (13)	31% (5)	13% (2)	0% (0)	13% (2)
Antihypertensives	89% (16)	28% (5)	0% (0)	0% (0)	22% (4)
Pain: nonopioid analgesics	100% (15)	0% (0)	0% (0)	0% (0)	0% (0)
Pain: opioid agonists	83% (10)	8% (1)	0% (0)	17% (2)	17% (2)
Seizure and neuropathic pain	57% (4)	43% (3)	14% (1)	14% (1)	43% (3)
Vitamins and herbal supplements	82% (103)	25% (31)	0% (0)	8% (10)	15% (19)
Other	95% (138)	12% (18)	1% (1)	5% (7)	21% (31)

Notes: GERD = gastroesophageal reflux disease; H2 = histamine 2; PPI = proton pump inhibitors. Percentages may add up to more than 100 because a response to one medication may fall into more than one category (e.g., participant states “I do not want to stop taking the medication because I have been taking it the last 10 years and it works well for me.” Response would be categorized as “pragmatic” and “appropriateness”).

Table 5.

Most Common Enablers by Medication Class

Medication category	Appropriateness	Influences	Process	Pragmatic	Dislike
Overall	67% (594)	23% (200)	23% (199)	19% (171)	5% (47)
Anticholinergic	73% (38)	25% (13)	19% (10)	21% (11)	2% (1)
Antipsychotic	64% (7)	27% (3)	18% (2)	45% (5)	0% (0)
Anxiety	66% (21)	25% (8)	28% (9)	28% (9)	3% (1)
Depression	59% (16)	26% (7)	33% (9)	19% (5)	4% (1)
Diabetes	67% (129)	24% (47)	26% (50)	16% (30)	5% (9)
GERD: PPI and H2 blocker	70% (33)	15% (7)	34% (16)	4% (2)	4% (2)
Antihypertensives	61% (46)	29% (22)	33% (25)	11% (8)	3% (2)
Pain: nonopioid analgesics	91% (29)	16% (5)	22% (7)	13% (4)	0% (0)
Pain: opioid agonists	79% (15)	11% (2)	16% (3)	11% (2)	11% (2)
Seizure and neuropathic pain	79% (11)	29% (4)	36% (5)	0% (0)	7% (1)
Vitamins and herbal supplements	68% (105)	19% (30)	15% (23)	25% (38)	10% (15)
Other	63% (142)	23% (52)	18% (40)	25% (57)	6% (13)

Notes: GERD = gastroesophageal reflux disease; H2 = histamine 2; PPI = proton pump inhibitors. Percentages may add up to more than 100 because a response to one medication may fall into more than one category (e.g., participant states “I would be willing to stop that medication if Dr. Smith approves and you monitor my symptoms and slowly work toward stopping.” Response would be categorized as “influences” and “process”).

Barriers by medication class

The frequency of specific barriers differed by medication indication. The only barrier to deprescribing antipsychotics was “fear” of cessation (100%). Patient/surrogate disagreement with the “appropriateness” of deprescribing medications for GERD (81%) was a more prominent barrier than “influences” (31%) or the “process” of discontinuation (13%). Disagreement with the “appropriateness” of deprescribing antihypertensives (89%) and medications for diabetes (83%) was cited by patients/surrogates more often than “influences” (28% and 19%, respectively) or “fear” of discontinuation (22% and 19%, respectively). Similar to other medication categories, patients/surrogates often expressed disagreement with the “appropriateness” of cessation of vitamins and herbal supplements (82%) and medications for depression (89%). “Influences” were barriers to deprescribing vitamins and herbal supplements (25%) more often than for medications for depression (4%). The “process” of discontinuing medications was rarely cited as a barrier to deprescribing, with the highest percentages found in seizure and neuropathic pain (14%) and GERD management (13%) medication categories.

Enablers by medication class

Enabler domains were also more prominent in certain medication categories. Patients/surrogates cited agreement with the “appropriateness” of deprescribing antipsychotic medications (64%) more often than “pragmatic” considerations related to its discontinuation (45%). The “appropriateness” of deprescribing medications for GERD (70%) was a more common enabler than the “process” of its discontinuation (34%) or “influences” (15%). Patients/surrogates agreed with the “appropriateness” of deprescribing for the majority of antihypertensives (61%), medications for diabetes (67%), vitamins and herbal supplements (68%), and medications for depression (59%) that were discussed during patient–clinician conversations. “Processes” were common enablers of deprescribing antihypertensives (33%) and seizure/neuropathic pain medications (36%); “influences” (29%) were identified as common deprescribing enablers in both of these medication categories as well (both 29%). “Pragmatic” considerations were most commonly cited as enablers of deprescribing medications for anxiety (28%) and vitamins and herbal supplements (25%). Rarely cited as an enabler of deprescribing across all medication classes, “dislike” of a medication was most commonly communicated during patient–clinician conversations about opioid analgesics (11%).

Discussion

Our study results mirrored previous literature findings of the pervasive use of PIMs among older adults with polypharmacy. Although the current deprescribing literature supports shared decision making, there is little research to guide clinicians on the actual implementation of deprescribing conversations (Turner et al., 2018). Clinicians often have difficulty communicating deprescribing recommendations to patients during routine clinical practice, citing a lack of self-efficacy as a barrier (Turner et al., 2018). It is reasonable to suspect that patient–clinician conversations may be even more challenging when patients have an accumulating number of medications and/or multiple chronic medical conditions. Although hospitalized older patients in this study had multimorbidity and were taking an average of 16.7 medications (indicative of hyperpolypharmacy), the high patient/surrogate agreement to study clinicians’ deprescribing recommendations (63%) demonstrates the importance of patient-centered conversations to support shared decision making and the use of a semistructured conversation tool to guide deprescribing conversations (Chock et al., 2021; Jansen et al., 2016; Turner et al., 2018; Weir et al., 2018).

Patient–Clinician Conversations

The semistructured conversation tool used in this study provided scaffolding for study clinicians while also allowing the conversations to be dynamic. This enabled the patient and/or surrogate to discuss the rationale(s) that motivated their agreement or disagreement with clinicians’ deprescribing recommendations and, in turn, allowed study clinicians to address potential barriers to deprescribing in real time. An integral part of our deprescribing tool was the set of questions we adapted from the BMQ (Svarstad et al., 1999). The standardized medication adherence and management questions and related tools yielded key information to help guide deprescribing decisions. Deprescribing infrastructure support for practicing clinicians is limited. The conversation tool and patient–clinician conversation process (Figure 1) used in this study can be used or modified by clinicians in their own practice environment to initiate and guide deprescribing conversations with older adult patients.

Deprescribing conversations in this study required a median of 30 min per patient and study clinicians discussed 82% of medications targeted for deprescribing during the prereview process. This indicates that the investment of time and resources to support these conversations in clinical settings is worthwhile for older adults with polypharmacy. Importantly, these conversations may be successfully led by team pharmacists or NPs, which may have important implications for increasing the feasibility of more widespread implementation and adoption in the hospital setting. Although only 53% of patients report comfort with pharmacist-initiated deprescribing (Hollingsworth et al., 2022), we found that patients were agreeable to deprescribing medications at a high percentage when conversations were led exclusively by a pharmacist or NP. The clinical roles and responsibilities of NPs and pharmacists are expanding (Funk & Weaver, 2018). These professions’ respective academic and training programs should include curricula to prepare graduates to collaboratively facilitate deprescribing processes and related patient conversations.

Barriers and Enablers

Patient-centered discussions allowed study clinicians to capture nuanced information that informed individual barriers and enablers to deprescribing. Knowledge of these barriers and enablers is valuable, as clinicians may use it to help frame ongoing deprescribing conversations with patients. It is valuable even prior to deprescribing (at initial prescription of a medication or during its continuation) as tailored patient education may make future deprescribing more amenable to a patient/surrogate. This study demonstrates the benefit of implementing an existing barriers and enablers framework during deprescribing conversations. Framing patients’ responses into themes maximizes clinicians’ understanding of factors that drive patients’ attitudes toward deprescribing and, in turn, may help improve the efficiency of future discussions and the sustainability of appropriate medication cessation. The patients in this study were hospitalized and were then transferred to PAC facilities for further care. This likely could have influenced both barriers and enablers to deprescribing. Further studies and deprescribing initiatives should examine the use of this barriers and enablers framework in other practice environments.

The “appropriateness” of medication use, or of its cessation, was the most common barrier to and enabler of deprescribing across all medication categories. As a barrier, patient/surrogate responses categorized under “appropriateness” were varied and individualized, making it challenging to understand as a whole. In many situations, patients/surrogates displayed a misunderstanding about a medication’s indication and, despite receiving tailored education from a study clinician, they were unwilling to engage in deprescribing it. Further research is needed to understand the complex relationship between patients’ expectations of a medication and clinician-delivered patient education, as many patients in our study agreed that deprescribing was “appropriate” after receiving tailored education from a study clinician.

Barriers to deprescribing

The “fear” of stopping a medication was predominantly expressed during discussions about antipsychotic medications and medications for seizures and neuropathic pain, two medical conditions for which patients/surrogates have little control. Patients’/surrogates’ reluctance to stop these medications is understandable, as initial use was likely associated with a significant event. This finding underscores the importance of provider empathy related to patients’/surrogates’ lived experiences and patient/surrogate trust in a clinician during deprescribing conversations (Zhang et al., 2018).

Enablers of deprescribing

“Processes” were common enablers of deprescribing. This reinforces the importance of a clinician explaining to the patient that deprescribing is a structured and planned process. Patients in this study were discharged to a PAC facility where they were followed by a clinician with whom study clinicians had regular contact. It is possible that being discharged to this type of monitored setting enabled the “process” of deprescribing.

Barriers and enablers by medication class

Overall, participants in this study were agreeable to the concept of deprescribing, but there remained variation in barriers to and enablers of deprescribing when specific medications were discussed as part of their overall medication regimen. This extends previous findings that the approach and content of patient–clinician deprescribing conversations should be adjusted based on the medication class (Turner et al., 2018). It also fills a knowledge gap, as most deprescribing interventions have been medication-specific. A noted limitation in the findings from Chock et al.’s (2021) recently published systematic review on deprescribing barriers and enablers is that many medication classes and diseases were not included in the review. The results of our study demonstrate that barriers to and enablers of deprescribing vary among patients and by medication class.

Commonly discussed medications

Patients in this study were least agreeable to deprescribing vitamins and herbal supplements, predominantly citing disagreement with the appropriateness of cessation and negative influences as barriers to deprescribing. The use of most vitamins and herbal supplements does not require a prescription; thus, these common OTC medications are widely available and may also not be considered during in-office medication reconciliation (Qato et al., 2016). These findings highlight the need for enhanced patient education on the relative benefits and harms of these types of medications, particularly among older patients with polypharmacy (Qato et al., 2016).

Patients’ high agreement to deprescribing medications for diabetes and for antihypertensives was predominantly driven by their agreement of its appropriateness. It is important to note that there are age-based prescribing and management guidelines for both of these medication categories, as the sequelae of overtreatment of these conditions in the older adult population can be significant, including exacerbation of geriatric syndromes (American Diabetes Association, 2020; Whelton et al., 2018).

Limitations

There are several notable limitations of this study. First, this study included older hospitalized adults with polypharmacy who required PAC at a facility. Enrollment in the larger trial was possibly influenced by patient or surrogate interest in deprescribing. Additionally, deprescribing conversations in this study were conducted within the hospital setting. Therefore, the rate of patient (or surrogate) agreement to deprescribing in the study may not generalize to other older adult patient populations or other clinical settings. We also did not monitor deprescribing conversations by routine providers; thus, we do not know if or how often these occurred and what influence these conversations may have had on patients’ deprescribing agreement. Finally, this study captured the patients’/surrogates’ reported agreement to deprescribing recommendations at the time of the transition from the hospital to a PAC facility; the extent to which this agreement translated to successful long-term deprescribing requires further investigation.

Conclusions and Implications

Older hospitalized patients (and/or their surrogates) discharging to PAC facilities are willing to have medications deprescribed when trained pharmacists or NPs initiate and facilitate patient-centered deprescribing conversations. Participants in this study expressed more agreement than disagreement with study clinicians’ deprescribing recommendations and the design of this study allowed us to elicit and address barriers and enablers to deprescribing in real time. These findings support an expanding role of pharmacists and NPs in the implementation of deprescribing interventions.

Funding

This work was supported by the National Institute on Aging at the National Institutes of Health (R01AG053264 to E. E. Vasilevskis and S. F. Simmons). The use of institutional data management system REDCap is supported by National Center for Advancing Translational Sciences (UL1TR000445).

Conflict of Interest

We have no potential financial or personal conflicts with any and all components of our research and content of the manuscript.